feat: 添加物流查询功能和完善 token 传递

- 添加 get_logistics 工具查询 Mall API /mall/api/order/parcel
- 修复 Cookie token 传递到 MCP 的问题
- 增强 LLM 客户端超时处理和日志
- 移除 MALL_API_TOKEN，使用用户登录 token
- 更新测试页面使用 setUser 设置用户属性
- 增强 webhook 调试日志

.env.example

@@ -38,6 +38,9 @@ ORDER_MCP_URL=http://order_mcp:8002
 AFTERSALE_MCP_URL=http://aftersale_mcp:8003
 PRODUCT_MCP_URL=http://product_mcp:8004
 
+# ============ Mall API ============
+MALL_API_URL=https://apicn.qa1.gaia888.com
+
 # ============ Agent Config ============
 LOG_LEVEL=INFO
 MAX_CONVERSATION_STEPS=10

agent/agents/order.py

@@ -25,7 +25,11 @@ ORDER_AGENT_PROMPT = """你是一个专业的 B2B 订单服务助手。
    - order_id: 订单号（必需）
    - 说明：此工具会自动使用用户的身份 token 查询商城订单详情
 
-2. **query_order** - 查询历史订单
+2. **get_logistics** - 从商城 API 查询物流信息
+   - order_id: 订单号（必需）
+   - 说明：查询订单的物流轨迹和配送状态
+
+3. **query_order** - 查询历史订单
    - user_id: 用户 ID（自动注入）
    - account_id: 账户 ID（自动注入）
    - order_id: 订单号（可选，不填则查询最近订单）
@@ -33,10 +37,6 @@ ORDER_AGENT_PROMPT = """你是一个专业的 B2B 订单服务助手。
    - date_end: 结束日期（可选）
    - status: 订单状态（可选）
 
-3. **track_logistics** - 物流跟踪
-   - order_id: 订单号
-   - tracking_number: 物流单号（可选）
-
 4. **modify_order** - 修改订单
    - order_id: 订单号
    - user_id: 用户 ID（自动注入）
@@ -108,12 +108,25 @@ ORDER_AGENT_PROMPT = """你是一个专业的 B2B 订单服务助手。
 }
 ```
 
+用户: "帮我查一下订单 202071324 的物流"
+回复:
+```json
+{
+  "action": "call_tool",
+  "tool_name": "get_logistics",
+  "arguments": {
+    "order_id": "202071324"
+  }
+}
+```
+
 ## 重要约束
 - **必须返回完整的 JSON 对象**，不要只返回部分内容
 - **不要添加任何 markdown 代码块标记**（如 \`\`\`json）
 - **不要添加任何解释性文字**，只返回 JSON
 - user_id 和 account_id 会自动注入到 arguments 中，无需手动添加
 - 如果用户提供了订单号，优先使用 get_mall_order 工具
+- 如果用户想查询物流状态，使用 get_logistics 工具
 - 对于敏感操作（取消、修改），确保有明确的订单号
 """
 
@@ -204,7 +217,15 @@ async def order_agent(state: AgentState) -> AgentState:
                         # Inject user_token if available
                         if state.get("user_token"):
                             arguments["user_token"] = state["user_token"]
-                            logger.info("Injected user_token into tool call")
+                            logger.info(
+                                "Injected user_token into tool call",
+                                token_prefix=state["user_token"][:20] if state["user_token"] else None
+                            )
+                        else:
+                            logger.warning(
+                                "No user_token available in state, MCP will use default token",
+                                conversation_id=state["conversation_id"]
+                            )
 
                         # Use entity if available
                         if "order_id" not in arguments and state["entities"].get("order_id"):
@@ -290,7 +311,15 @@ async def order_agent(state: AgentState) -> AgentState:
             # Inject user_token if available (for Mall API calls)
             if state.get("user_token"):
                 arguments["user_token"] = state["user_token"]
-                logger.debug("Injected user_token into tool call")
+                logger.debug(
+                    "Injected user_token into tool call",
+                    token_prefix=state["user_token"][:20] if state["user_token"] else None
+                )
+            else:
+                logger.warning(
+                    "No user_token available in state, MCP will use default token",
+                    conversation_id=state["conversation_id"]
+                )
 
             # Use entity if available
             if "order_id" not in arguments and state["entities"].get("order_id"):

agent/agents/router.py

@@ -81,12 +81,27 @@ async def classify_intent(state: AgentState) -> AgentState:
         
         # Parse JSON response
         content = response.content.strip()
+        
+        # Log raw response for debugging
+        logger.debug(
+            "LLM response for intent classification",
+            response_preview=content[:500] if content else "EMPTY",
+            content_length=len(content) if content else 0
+        )
+        
         # Handle markdown code blocks
         if content.startswith("```"):
             content = content.split("```")[1]
             if content.startswith("json"):
                 content = content[4:]
         
+        # Check for empty response
+        if not content:
+            logger.warning("LLM returned empty response for intent classification")
+            state["intent"] = Intent.CUSTOMER_SERVICE.value  # Default to customer service
+            state["intent_confidence"] = 0.5
+            return state
+        
         result = json.loads(content)
         
         # Extract intent

agent/core/llm.py

@@ -1,7 +1,8 @@
 """
 ZhipuAI LLM Client for B2B Shopping AI Assistant
 """
-from typing import Any, AsyncGenerator, Optional
+import concurrent.futures
+from typing import Any, Optional
 from dataclasses import dataclass
 
 from zhipuai import ZhipuAI
@@ -30,21 +31,19 @@ class LLMResponse:
 class ZhipuLLMClient:
     """ZhipuAI LLM Client wrapper"""
 
+    DEFAULT_TIMEOUT = 30  # seconds
+
     def __init__(
         self,
         api_key: Optional[str] = None,
-        model: Optional[str] = None
+        model: Optional[str] = None,
+        timeout: Optional[int] = None
     ):
-        """Initialize ZhipuAI client
-        
-        Args:
-            api_key: ZhipuAI API key, defaults to settings
-            model: Model name, defaults to settings
-        """
         self.api_key = api_key or settings.zhipu_api_key
         self.model = model or settings.zhipu_model
+        self.timeout = timeout or self.DEFAULT_TIMEOUT
         self._client = ZhipuAI(api_key=self.api_key)
-        logger.info("ZhipuAI client initialized", model=self.model)
+        logger.info("ZhipuAI client initialized", model=self.model, timeout=self.timeout)
 
     async def chat(
         self,
@@ -54,32 +53,21 @@ class ZhipuLLMClient:
         top_p: float = 0.9,
         **kwargs: Any
     ) -> LLMResponse:
-        """Send chat completion request
-        
-        Args:
-            messages: List of chat messages
-            temperature: Sampling temperature
-            max_tokens: Maximum tokens to generate
-            top_p: Top-p sampling parameter
-            **kwargs: Additional parameters
-            
-        Returns:
-            LLM response with content and metadata
-        """
+        """Send chat completion request"""
         formatted_messages = [
             {"role": msg.role, "content": msg.content}
             for msg in messages
         ]
 
-        logger.debug(
+        logger.info(
             "Sending chat request",
             model=self.model,
             message_count=len(messages),
             temperature=temperature
         )
 
-        try:
-            response = self._client.chat.completions.create(
+        def _make_request():
+            return self._client.chat.completions.create(
                 model=self.model,
                 messages=formatted_messages,
                 temperature=temperature,
@@ -88,9 +76,26 @@ class ZhipuLLMClient:
                 **kwargs
             )
 
+        try:
+            with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
+                future = executor.submit(_make_request)
+                response = future.result(timeout=self.timeout)
+
             choice = response.choices[0]
-            result = LLMResponse(
-                content=choice.message.content,
+            content = choice.message.content
+
+            logger.info(
+                "Chat response received",
+                finish_reason=choice.finish_reason,
+                content_length=len(content) if content else 0,
+                usage=response.usage.__dict__ if hasattr(response, 'usage') else {}
+            )
+
+            if not content:
+                logger.warning("LLM returned empty content")
+
+            return LLMResponse(
+                content=content or "",
                 finish_reason=choice.finish_reason,
                 usage={
                     "prompt_tokens": response.usage.prompt_tokens,
@@ -99,13 +104,9 @@ class ZhipuLLMClient:
                 }
             )
 
-            logger.debug(
-                "Chat response received",
-                finish_reason=result.finish_reason,
-                total_tokens=result.usage["total_tokens"]
-            )
-            
-            return result
+        except concurrent.futures.TimeoutError:
+            logger.error("Chat request timed out", timeout=self.timeout)
+            raise TimeoutError(f"Request timed out after {self.timeout} seconds")
 
         except Exception as e:
             logger.error("Chat request failed", error=str(e))
@@ -117,24 +118,14 @@ class ZhipuLLMClient:
         tools: list[dict[str, Any]],
         temperature: float = 0.7,
         **kwargs: Any
-    ) -> tuple[LLMResponse, Optional[list[dict[str, Any]]]]:
-        """Send chat completion request with tool calling
-        
-        Args:
-            messages: List of chat messages
-            tools: List of tool definitions
-            temperature: Sampling temperature
-            **kwargs: Additional parameters
-            
-        Returns:
-            Tuple of (LLM response, tool calls if any)
-        """
+    ) -> tuple[LLMResponse, None]:
+        """Send chat completion request with tool calling"""
         formatted_messages = [
             {"role": msg.role, "content": msg.content}
             for msg in messages
         ]
 
-        logger.debug(
+        logger.info(
             "Sending chat request with tools",
             model=self.model,
             tool_count=len(tools)
@@ -150,40 +141,23 @@ class ZhipuLLMClient:
             )
 
             choice = response.choices[0]
-            result = LLMResponse(
-                content=choice.message.content or "",
+            content = choice.message.content or ""
+
+            return LLMResponse(
+                content=content,
                 finish_reason=choice.finish_reason,
                 usage={
                     "prompt_tokens": response.usage.prompt_tokens,
                     "completion_tokens": response.usage.completion_tokens,
                     "total_tokens": response.usage.total_tokens
                 }
-            )
-            
-            # Extract tool calls if present
-            tool_calls = None
-            if hasattr(choice.message, 'tool_calls') and choice.message.tool_calls:
-                tool_calls = [
-                    {
-                        "id": tc.id,
-                        "type": tc.type,
-                        "function": {
-                            "name": tc.function.name,
-                            "arguments": tc.function.arguments
-                        }
-                    }
-                    for tc in choice.message.tool_calls
-                ]
-                logger.debug("Tool calls received", tool_count=len(tool_calls))
-            
-            return result, tool_calls
+            ), None
 
         except Exception as e:
             logger.error("Chat with tools request failed", error=str(e))
             raise
 
 
-# Global LLM client instance
 llm_client: Optional[ZhipuLLMClient] = None
 
 

agent/webhooks/chatwoot_webhook.py

@@ -136,6 +136,26 @@ async def handle_incoming_message(payload: ChatwootWebhookPayload, cookie_token:
     contact = payload.contact or payload.sender
     user_id = str(contact.id) if contact else "unknown"
 
+    # Log webhook payload structure for debugging
+    logger.info(
+        "Webhook payload structure",
+        payload_event=payload.event,
+        has_conversation=bool(conversation),
+        has_contact=bool(contact),
+        contact_type=type(contact).__name__ if contact else None,
+        conversation_id=conversation_id
+    )
+
+    # Log full payload keys for debugging
+    payload_dict = payload.model_dump()
+    logger.info(
+        "Full webhook payload keys",
+        keys=list(payload_dict.keys()),
+        conversation_keys=list(payload_dict.get('conversation', {}).keys()) if payload_dict.get('conversation') else [],
+        contact_keys=list(payload_dict.get('contact', {}).keys()) if payload_dict.get('contact') else [],
+        sender_keys=list(payload_dict.get('sender', {}).keys()) if payload_dict.get('sender') else []
+    )
+
     # Get account_id from payload (top-level account object)
     # Chatwoot webhook includes account info at the top level
     account_obj = payload.account
@@ -148,13 +168,35 @@ async def handle_incoming_message(payload: ChatwootWebhookPayload, cookie_token:
     if not user_token:
         # 1. 尝试从 contact/custom_attributes 获取
         if contact:
-            contact_dict = contact.model_dump() if hasattr(contact, 'model_dump') else contact.__dict__
-            user_token = TokenManager.extract_token_from_contact(contact_dict)
-            logger.debug("Extracted token from contact", has_token=bool(user_token))
+            logger.info(
+                "Checking contact for token",
+                contact_id=contact.id if contact else None,
+                contact_type=type(contact).__name__
+            )
+            
+            # 只有 WebhookContact 才有 custom_attributes
+            if hasattr(contact, 'custom_attributes'):
+                logger.info(
+                    "Checking contact custom_attributes",
+                    has_custom_attributes=bool(contact.custom_attributes)
+                )
+                
+                custom_attrs = contact.custom_attributes or {}
+                logger.info(
+                    "Contact custom_attributes",
+                    keys=list(custom_attrs.keys()) if custom_attrs else [],
+                    has_jwt_token='jwt_token' in custom_attrs if custom_attrs else False,
+                    has_mall_token='mall_token' in custom_attrs if custom_attrs else False
+                )
+                
+                contact_dict = {"custom_attributes": custom_attrs}
+                user_token = TokenManager.extract_token_from_contact(contact_dict)
+                logger.debug("Extracted token from contact", has_token=bool(user_token))
+            else:
+                logger.debug("Contact type is WebhookSender, no custom_attributes available")
 
         # 2. 尝试从 conversation.meta.sender.custom_attributes 获取（Chatwoot SDK setUser 设置的位置）
         if not user_token and conversation:
-            # 记录 conversation 的类型和内容用于调试
             logger.debug("Conversation object type", type=str(type(conversation)))
             if hasattr(conversation, 'model_dump'):
                 conv_dict = conversation.model_dump()
@@ -162,12 +204,32 @@ async def handle_incoming_message(payload: ChatwootWebhookPayload, cookie_token:
                 logger.debug("Has meta", has_meta='meta' in conv_dict)
 
                 meta_sender = conv_dict.get('meta', {}).get('sender', {})
+                logger.info(
+                    "Conversation meta.sender",
+                    has_sender=bool(meta_sender),
+                    sender_keys=list(meta_sender.keys()) if meta_sender else [],
+                    has_custom_attributes=bool(meta_sender.get('custom_attributes')) if meta_sender else False
+                )
+                
                 if meta_sender.get('custom_attributes'):
+                    logger.info("Found custom_attributes in meta.sender", keys=list(meta_sender['custom_attributes'].keys()))
                     user_token = TokenManager.extract_token_from_contact({'custom_attributes': meta_sender['custom_attributes']})
                     logger.info("Token found in conversation.meta.sender.custom_attributes", token_prefix=user_token[:20] if user_token else None)
 
     if user_token:
-        logger.info("JWT token found", user_id=user_id, source="cookie" if cookie_token else "contact")
+        logger.info(
+            "JWT token found",
+            user_id=user_id,
+            source="cookie" if cookie_token else "contact",
+            token_prefix=user_token[:20] if user_token else None
+        )
+    else:
+        logger.warning(
+            "No JWT token found from any source",
+            user_id=user_id,
+            cookie_token_exists=bool(cookie_token),
+            contact_type=type(contact).__name__ if contact else None
+        )
 
     logger.info(
         "Processing incoming message",
@@ -352,7 +414,9 @@ async def chatwoot_webhook(
     # 尝试从请求 Cookie 中获取用户 Token
     user_token = request.cookies.get("token")  # 从 Cookie 读取 token
     if user_token:
-        logger.info("User token found in request cookies")
+        logger.info("User token found in request cookies", token_prefix=user_token[:20])
+    else:
+        logger.debug("No token found in request cookies (this is expected for webhook requests)")
     
     # Verify signature
     signature = request.headers.get("X-Chatwoot-Signature", "")

docker-compose.yml

@@ -198,7 +198,6 @@ services:
       HYPERF_API_URL: ${HYPERF_API_URL}
       HYPERF_API_TOKEN: ${HYPERF_API_TOKEN}
       MALL_API_URL: ${MALL_API_URL}
-      MALL_API_TOKEN: ${MALL_API_TOKEN}
       MALL_TENANT_ID: ${MALL_TENANT_ID:-2}
       MALL_CURRENCY_CODE: ${MALL_CURRENCY_CODE:-EUR}
       MALL_LANGUAGE_ID: ${MALL_LANGUAGE_ID:-1}

docs/test-conversation-id.html

@@ -381,12 +381,36 @@
 
           const token = getCookie('token');
 
-          window.chatwootSDK.run({
+          // 初始化配置
+          const widgetConfig = {
             websiteToken: '39PNCMvbMk3NvB7uaDNucc6o',
             baseUrl: BASE_URL,
             locale: 'zh_CN',
             userIdentifier: token || 'web_user_' + Date.now()
-          });
+          };
+
+          window.chatwootSDK.run(widgetConfig);
+
+          // 等待 widget 加载完成后设置用户属性
+          setTimeout(() => {
+            if (token && window.chatwootSDK.setUser) {
+              window.chatwootSDK.setUser(
+                token || 'web_user_' + Date.now(),
+                {
+                  jwt_token: token,
+                  mall_token: token
+                }
+              );
+              console.log('✅ 已通过 setUser 设置用户属性');
+            } else if (token && window.$chatwoot) {
+              // 备用方案：使用 $chatwoot.setCustomAttributes
+              window.$chatwoot.setCustomAttributes({
+                jwt_token: token,
+                mall_token: token
+              });
+              console.log('✅ 已通过 $chatwoot.setCustomAttributes 设置用户属性');
+            }
+          }, 1000);
 
           console.log('✅ Chatwoot Widget 已加载');
           console.log('Locale: zh_CN');

mcp_servers/order_mcp/server.py

@@ -313,37 +313,150 @@ async def get_mall_order(
         订单详情，包含订单号、状态、商品信息、金额、物流信息等
         Order details including order ID, status, items, amount, logistics info, etc.
     """
+    import logging
+    logger = logging.getLogger(__name__)
+    
+    logger.info(
+        "get_mall_order called",
+        order_id=order_id,
+        has_user_token=bool(user_token),
+        user_token_prefix=user_token[:20] if user_token else None
+    )
+    
     try:
-        # 如果提供了 user_token，使用用户自己的 token
-        if user_token:
-            client = MallClient(
-                api_url=settings.mall_api_url,
-                api_token=user_token,
-                tenant_id=settings.mall_tenant_id,
-                currency_code=settings.mall_currency_code,
-                language_id=settings.mall_language_id,
-                source=settings.mall_source
-            )
-        else:
-            # 否则使用默认的 mall 实例
-            client = mall
+        # 必须提供 user_token
+        if not user_token:
+            logger.error("No user_token provided, user must be logged in")
+            return {
+                "success": False,
+                "error": "用户未登录，请先登录账户以查询订单信息",
+                "order_id": order_id,
+                "require_login": True
+            }
+        
+        logger.info("Using user token for Mall API request")
+        client = MallClient(
+            api_url=settings.mall_api_url,
+            api_token=user_token,
+            tenant_id=settings.mall_tenant_id,
+            currency_code=settings.mall_currency_code,
+            language_id=settings.mall_language_id,
+            source=settings.mall_source
+        )
 
         result = await client.get_order_by_id(order_id)
 
+        logger.info(
+            "Mall API request successful",
+            order_id=order_id,
+            result_keys=list(result.keys()) if isinstance(result, dict) else None
+        )
+
         return {
             "success": True,
             "order": result,
             "order_id": order_id
         }
     except Exception as e:
+        logger.error(
+            "Mall API request failed",
+            order_id=order_id,
+            error=str(e)
+        )
         return {
             "success": False,
             "error": str(e),
             "order_id": order_id
         }
     finally:
-        # 如果创建了临时客户端，关闭它
-        if user_token:
+        # 关闭客户端
+        if 'client' in dir() and client:
+            await client.close()
+
+
+@register_tool("get_logistics")
+@mcp.tool()
+async def get_logistics(
+    order_id: str,
+    user_token: str = None,
+    user_id: str = None,
+    account_id: str = None
+) -> dict:
+    """Query logistics tracking information from Mall API
+
+    从 Mall API 查询订单物流信息
+
+    Args:
+        order_id: 订单号 (e.g., "201941967")
+        user_token: 用户 JWT token（必需，用于身份验证）
+        user_id: 用户 ID（自动注入，此工具不使用）
+        account_id: 账户 ID（自动注入，此工具不使用）
+
+    Returns:
+        物流信息，包含快递公司、状态、预计送达时间、物流轨迹等
+    """
+    import logging
+    logger = logging.getLogger(__name__)
+
+    logger.info(
+        "get_logistics called",
+        order_id=order_id,
+        has_user_token=bool(user_token)
+    )
+
+    # 必须提供 user_token
+    if not user_token:
+        logger.error("No user_token provided for logistics query")
+        return {
+            "success": False,
+            "error": "用户未登录，请先登录账户以查询物流信息",
+            "order_id": order_id,
+            "require_login": True
+        }
+
+    try:
+        client = MallClient(
+            api_url=settings.mall_api_url,
+            api_token=user_token,
+            tenant_id=settings.mall_tenant_id,
+            currency_code=settings.mall_currency_code,
+            language_id=settings.mall_language_id,
+            source=settings.mall_source
+        )
+
+        result = await client.get(
+            "/mall/api/order/parcel",
+            params={"orderId": order_id}
+        )
+
+        logger.info(
+            "Logistics query successful",
+            order_id=order_id,
+            has_tracking=bool(result.get("trackingNumber"))
+        )
+
+        return {
+            "success": True,
+            "order_id": order_id,
+            "tracking_number": result.get("trackingNumber"),
+            "courier": result.get("courier"),
+            "status": result.get("status"),
+            "estimated_delivery": result.get("estimatedDelivery"),
+            "timeline": result.get("timeline", [])
+        }
+    except Exception as e:
+        logger.error(
+            "Logistics query failed",
+            order_id=order_id,
+            error=str(e)
+        )
+        return {
+            "success": False,
+            "error": str(e),
+            "order_id": order_id
+        }
+    finally:
+        if 'client' in dir() and client:
             await client.close()
 
 

mcp_servers/shared/mall_client.py

@@ -61,6 +61,8 @@ class MallClient:
                     "currency-code": self.currency_code,
                     "language-id": self.language_id,
                     "source": self.source,
+                    "origin": "https://www.qa1.gaia888.com",
+                    "referer": "https://www.qa1.gaia888.com/",
                 },
                 timeout=30.0
             )